Question that shows up every week on Valorant forums: does Vanguard catch people using voice changers? Short answer: no. Long answer explains why — and shows the setup that works for whoever wants to swap voices in Valorant without turning into a support ticket case.
What Vanguard actually watches
Vanguard is a kernel-mode anti-cheat. It loads before Windows finishes booting and keeps a supervisor driver always active. Its scope covers:
valorant.exeprocess memory- Suspicious kernel drivers (low-level cheat injectors)
- Hooks in game DLLs
- Runtime CPU instruction modifications
What it doesn’t monitor: the Windows audio subsystem, sound card drivers, capture software like OBS, user-mode audio processing software. VoxBooster operates 100% in that area. Vanguard has no way to see it, and even if it did, it wouldn’t classify it as cheating — Riot bans for aim, walls and radar leaks, not for how you sound.
Straight setup
- Install VoxBooster.
- Open it, log in, pick the tab that makes sense (Voice Clone if you want to sound like another person, Voice Effects if you just want a comic character).
- Flip “Real-time” on.
- Open Valorant, go to Settings → Audio → Voice.
- Confirm the Input Device is your real mic (not VoxBooster, not any virtual device).
- Play.
Done. No secret step, no driver to install, no PC reboot. Valorant captures your already-transformed voice because VoxBooster intercepts the audio before Windows hands it to any application.
Specific settings that matter
Valorant mic volume: keep it between 70-85%. Above that, the game applies its own gain that saturates the already-modified signal, causing distortion. Below 60%, the team can’t hear you properly.
Push to Talk: always use it. Voice activation in Valorant detects mouse clicks and keyboard taps as voice, and the clone ends up processing noise for no reason. PTT bound on the mouse (M4 or M5) is what most pros run, and it works perfectly with VoxBooster.
Valorant’s noise filter: disable it. VoxBooster already has its own noise suppression that runs before the transformation. Valorant’s filter applied on top of the transformed voice produces a horrible robotic artifact. In Audio → Voice, set “Echo / Noise Cancellation” to Off.
Practical latency for ranked
On a mid-range PC:
- Voice Effects (any preset): ~5ms
- Neural Voice Clone: ~480ms
- Voice Clone low-latency mode: ~250ms
For ranked, 5ms is invisible — you use the effect without noticing the delay. For the neural clone, 250ms is still fine for calls like “rotating to B” but hurts if you’re the IGL calling reads in real time (“flash now, push push”). Most elo-ranked players use effects; cloning stays for unranked, deathmatch, replay, and custom modes.
Combos that look good in VAL
Each agent has a vibe and some voices click:
- Jett main — young female voice, agile, slightly pitched up
- Cypher main — voice with light echo and mysterious tone
- Sage main — soft, calm female voice
- Brimstone main — deep military commander voice
- Iso main — low, paced, intimidating voice
No rule — the fun is specifically breaking stereotypes (Sage with a demon voice is comedy gold). Depends on stream type and squad vibe.
Soundboard for crisp clips
Global bind for:
- “Accept duel” Iso sample every time you duel
- Victory horn for last-round defuse
- “Ezi” from TenZ for clips
- “ROUND OVER” narrator voice ripped from the game itself
Hotkey works with Valorant in fullscreen because VoxBooster uses global Windows hooks, doesn’t depend on window focus. Hit the key mid-round, sound plays for the team, you keep playing.